Intermittently getting an "unsupported voice" error when using the Ash voice

API
1

I’m working on a project to generate mp3 files for a Bible translation. I’m submitting a chapter at a time. When I use the “ash” voice I intermittently get this error:

Response: { “error”: { “message”: “[{‘type’: ‘enum’, ‘loc’: (‘body’, ‘voice’), ‘msg’: "Input should be ‘nova’, ‘shimmer’, ‘echo’, ‘onyx’, ‘fable’ or ‘alloy’", ‘input’: ‘ash’, ‘ctx’: {‘expected’: "‘nova’, ‘shimmer’, ‘echo’, ‘onyx’, ‘fable’ or ‘alloy’"}}]”, “type”: “invalid_request_error”, “param”: null, “code”: null } }

It looks like an error I should get before the three new voices (ash, coral, and sage) were made available a few months ago. I can run it x number of times successfully, but then the error happens.

Edit to add that this never happens with any of the six original voices. It’s only with the three new ones.

Thoughts? Ideas?
Thanks!

2

So nobody else has experienced this problem?

For the record, I just now experienced it in the TTS Playground. I’m using the Ash voice, testing things out, seeing where phonetic spelling for some words works better, and I got the same error as above: “ash” is unexpected input. I clicked “Generate” again, and it worked.

I’m no expert, but it almost seems like when the changes were rolled out to handle the three new voices, somehow a server missed the update, and when you submit a request to generate speech, if you’re unlucky and get sent to that server, you get the error.

On my programming calls to the API, if I submit a request and get the error, I simply resubmit it (up to three times). So far so good.

IMO, the fact that I got the error in the TTS Playground demonstrates that this really is a problem.

Thanks

3

@rostwoods Thanks for reporting this, and sorry for the trouble. It’s now fixed!

1 Like
4

Thanks much!! So I wasn’t nuts after all… :sunglasses: